Wan 3.0 香港實測攻略：阿里影片規格、成本與提示詞指南 (2026)

提示詞結構 = [整體描述 (Overall description)]
           + [參考素材標記 (Reference material citation): Image N / Video N / Audio N]
           + [分鏡 Shot N (起始-結束秒數): 主體 + 場景 + 動作 + 鏡頭控制]
           + [角色對白 (Dialogue): Character says: "Line"]
           + [音效 / 背景音樂 (Sound effect / BGM)]
           + [畫面風格 / 氛圍 (Style / Mood)]
           + [負向提示詞 (Negative prompt list)]

Period drama sequence in an 18th-century library. Lord Malvern (referencing Image 1) confronts Julian (referencing Image 2) across an oak table. 15 seconds duration.

Shot 1 (00:00-00:05): Wide shot, 28mm lens. Lord Malvern stands by hearth holding an unsealed envelope.

Shot 2 (00:05-00:10): Medium shot, 50mm lens. Julian enters through oak doors. Lord Malvern speaks with quiet tension: "You should not have returned to the estate."

Shot 3 (00:10-00:15): Close-up, 85mm lens. Julian replies: "The inheritance belongs to the county now."

Sound design: Crackling fireplace wood, heavy window rain, clear spoken dialogue. Mood: Somber British period drama, warm amber key light, dark shadow falloff.

A 15-second wildlife tracking sequence across savanna grasslands during golden hour.

Shot 1 (00:00-00:05): Extreme wide shot. Camera glides forward across savanna while tilting upward as a cheetah enters frame left on the path of a fleeing antelope.

Shot 2 (00:05-00:10): Low tracking shot parallel to the cheetah on the right. Grassy background sweeps past rapidly while subject stays left of center.

Shot 3 (00:10-00:15): Forward push-in at ground level, narrowing focus as the cheetah extends forelimbs toward the antelope's hindquarters.

Sound design: Crescendoing orchestral score, rhythmic drum cadence, paw thuds on dry soil, rushing wind. Final second cuts to isolated breathing.

A 15-second warm-toned commercial brand film for an artisanal café, referencing barista in Image 1 and color palette in Image 2.

Shot 1 (00:00-00:05): Macro close-up shot, 100mm lens. Steaming water pours from brass kettle onto ground coffee in ceramic dripper. Water droplets catch morning window light.

Shot 2 (00:05-00:10): Medium tracking shot along a timber counter. A barista in an indigo apron slides a warm ceramic mug toward an unseen customer beside a sunlit window.

Shot 3 (00:10-00:15): Close-up portrait, 85mm lens. A smiling patron wraps hands around the ceramic cup, takes an appreciative sip, and looks toward the autumn street.

Sound design: Gentle water pouring hiss, ceramic mug clink on wood, acoustic guitar melody, ambient café murmurs. Mood: Warm golden tones.

Reference-to-video motion transfer: Replicate camera movement, sword strike trajectory, and footwork from Video 1, casting armored warrior from Image 1 as performer.

Action description: Armored warrior executes 360-degree spinning broadsword parry against two incoming spear attacks in stone courtyard. Maintain velocity curves and low whip-pan from Video 1.

Shot parameters: 10 seconds duration, 1080p resolution, 16:9 widescreen ratio. Armor engraving and face scars match Image 1.

Sound design: Heavy metal blade clashes, rain on steel plate, combat vocal grunts. Mood: Gritty dark fantasy, desaturated slate tones with bright weapon sparks.

Vertical 9:16 short drama scene in a corporate boardroom at night. Elena (referencing Image 1) confronts Marcus (referencing Image 2) across a glass conference table. 10 seconds duration.

Shot 1 (00:00-00:05): Medium close-up, vertical framing, 50mm lens. Elena stands under cool fluorescent lights, drops a signed contract onto the glass, and says firmly: "The board voted ten minutes ago."

Shot 2 (00:05-00:10): Close-up, 85mm lens. Marcus leans back into his leather chair, stares up with narrowed eyes, and answers quietly: "They don't own the voting shares, Elena."

Sound design: Sharp paper slap against glass table, low hum of office air conditioning, distant city sirens, clear synchronized dialogue. Mood: High-stakes corporate thriller, cool blue rim light with sharp shadows.